On 30.09.20 at 11:23 Jiri Denemark wrote: > On Mon, Sep 28, 2020 at 13:20:06 +0200, Markus Schade wrote: >> Okay, that did not go as expected. Anyway > > I'm not sure what did not go as expected... I had meant to sent the cover letter with the patch ;-) >> as well as updated cpuid information from a Ryzen 9 3900 system. > > These changes should be dropped. We already have this model covered and > regenerating the data files for existing CPUs just causes unnecessary > and often unrelated changes. These updated data files are required because libvirt will now try to use the better matching EPYC-Rome model for Ryzen 9 tests. So leaving them out, will break tests. Just changing the model to EPYC-Rome in the Ryzen 9 tests files, will also breaks tests. Hence I re-ran the gather/parse on a Ryzen 9 system to get the updated cpuid information. >> Since the EPYC 7502 and the Ryzen 9 are pretty much identical, the >> latter could be dropped as there is no corresponding qemu model. > > While we don't have a special model Ryzen 9, we still want to have it > covered in the test suit to check we assign the appropriate model to it. Got it. Markus
Attachment:
signature.asc
Description: OpenPGP digital signature
